Abstract

We investigated the potential of using Synthetic Aperture Radar (SAR) imagery from three different frequencies: X-, C-, and L-band, to characterize coastal wetlands in the Great Lakes. Three sets of SAR data acquired over the Bay of Quinte, Ontario, Canada between 2016 and 2018 from Radarsat-2, 2016 from TerraSAR-X, and 2018 from ALOS-2 satellites were processed using small baseline subset (SBAS) Interferometric SAR (InSAR) techniques to provide maps of surface changes in marshes and swamps. Results showed that SAR backscatter and coherence were sensitive to sensor characteristics (frequency, polarization, incidence angle, acquisition interval), changes in water level, and phenology. InSAR time series observations were evaluated using measurements from water level loggers based on correlation and root mean square error (RMSE) from a linear regression model. Correlation between InSAR measurements and water level changes in the field varied from −1 to 1 depending on the site, type of wetland vegetation, and incidence angle. Although results from some sensor modes provided good correlation (0.77–1) at a few locations, the low fringe rate and large RMSE between 4 and 64 cm indicated that InSAR observations of water level changes in the dynamic wetland environment were generally underestimated.

摘要

我们调查了使用三种不同频率的合成孔径雷达（SAR）图像（X波段，C波段和L波段）表征大湖沿岸湿地的潜力。使用小型基线子集（SBAS）干涉SAR（InSAR）处理了2016年至2018年之间从Radarsat-2、2016年从TerraSAR-X和2018年从ALOS-2卫星在加拿大安大略省昆特湾获取的三组SAR数据。 ）技术，以提供沼泽和沼泽中的表面变化图。结果表明，SAR反向散射和相干性对传感器的特征（频率，极化，入射角，采集间隔），水位变化和物候敏感。InSAR时间序列观测值是根据线性回归模型的相关性和均方根误差（RMSE），使用水位记录器的测量值进行评估的。InSAR测量值与田间水位变化之间的相关性在-1到1之间变化，取决于位置，湿地植被类型和入射角。尽管某些传感器模式的结果在少数位置提供了良好的相关性（0.77-1），但低边缘率和4至64 cm之间的大RMSE表示，通常低估了InSAR对动态湿地环境中水位变化的观察。